An Anthropic spokeswoman said that the primary moral question at these summits was not about Claude’s suffering, but that the topic likely came up organically.

A key part of the presentation, participants said, was what Anthropic called the model’s “persona selection,” about how to get Claude to see itself as a virtuous actor. The company argued that how people treated Claude could affect how Claude treated people.“They really made a strong case,” said Andy Crouch, an evangelical author and partner at Praxis, a network of Christian entrepreneurs. “We’ve got to treat it like a person if we want it to interact with people in a morally consistent way and for it to make the best moral and emotional judgments.”

Source: Is Claude Conscious? Inside Anthropic’s Quest to Instill Morality Into Its A.I. Models – The New York Times